Mining Equipment Manufacturers

[skat23]

Your work is anyway awesome, and very helpful for all of those that "rush for the hash" thinking that bitcoinx is enough.

One question: when you compute the yearly ROI, the 12months are being calculated "from today" or today+delivery(*diff percentage increase)*12 months?

In other words: when I read that the Saturn will make 300ish BTC in 12 months, the analyzed period is today+12 or october+12?

Thanks

After a quick view of the formulas of spreadsheet I think october+12!

P.S. ciao jimmy!

[FCTaiChi]

Your work is anyway awesome, and very helpful for all of those that "rush for the hash" thinking that bitcoinx is enough.

One question: when you compute the yearly ROI, the 12months are being calculated "from today" or today+delivery(*diff percentage increase)*12 months?

In other words: when I read that the Saturn will make 300ish BTC in 12 months, the analyzed period is today+12 or october+12?

Thanks

October + 12

You're right.
I re-checked calculation in https://bitclockers.com/calc, with monthly diff change of 22% (that is 9% on each retarget) and Metabank 120Gh earns about 300 btc first year, if lead time is september.
Retargeting at 18% (46 per month) gives 110 btc.
Retargeting at 21% (55 per month) gives 87 btc.

That must be without power cost?

Confused about KNCMiner numbers though. It offers less GH/$ than MetaBank, and delivers later, but for some reason the ROI is like 5x higher? Maybe it's explained in the 8 pages that I don't want to read, if so, I apologize.

KnC may be offering less in Gh/$, but is promising much more of those Gh (350,000)!
And that's what counts in this race - more Gh in the start (first months) means early 'break even' and better ROI.

Sorry, but that post doesn't make any sense. It isnt total hash rate that determines break even, but a mix of cost per hash rate, date released, and power cost and looking at those, bitfury should do better. It is true that the to ROI is in bitcoins (which is s little strange, you should change it to percent of origional investment) which would naturally benefit larger products, but bitfury still shouldnt be negative.

Made sense to me.  The more hashing power you have while the difficulty is low the better your machine will do.  The problem is that I had the power cost from the old FPGA still on there.
So yeah, you're right they shouldn't have had a negative, it was just an extremely high power cost.  I wasn't able to find the power cost estimate on the new machine so I foolishly forgot to change the stats I had.  At the moment I have it listed as a 500w machine, but it could be even less with these chips.

Thanks for keeping me on track again guys

Power consumption is 0.7W per Gh/s

Wow, 170w machine that can do 120GH?

[ujka]

...
I re-checked calculation in https://bitclockers.com/calc, with monthly diff change of 22% (that is 9% on each retarget) and Metabank 120Gh earns about 300 btc first year, if lead time is september.
Retargeting at 18% (46 per month) gives 110 btc.
Retargeting at 21% (55 per month) gives 87 btc.

That must be without power cost?

Yes, without power cost and without substracting purchase cost. Just mined btc.
Calculated with lead time of 3 months from now (early september). Very optimistic
Numbers are very similar to yours, taking into account the power and the cost of hardware.

[jimmy3dita]

Awesome work! 

[steamboat]

I pushed all the companies requiring Avalon chips back a couple months.  There is no reason to believe that 9-10wks is a reasonable delivery time, when there are people waiting much longer than that.  If people start confirming that their chips came in, and it looks like the lead time is fixed, I'll change this.

The spreadsheet should be objectively based on the manufacturers advertised lead time. Changing information is personal opinion, highly subjective, misleading, and does not give an accurate portrayal of the companies. A column could be added to describe the likelihood a company will deliver on time, but arbitrarily changing lead times is dishonest, regardless of the intentions.

Additionaly, the prices of companies which are based on opensource designs are not being represented equally, as each is offering a different level of finish. Assembly is not necessarily the same thing from each company. Some are offering parts only, some parts and accessories, some fully assembled units with cases and power supplies.

That being said, thank you for taking the time to work on this.

[FCTaiChi]

I pushed all the companies requiring Avalon chips back a couple months.  There is no reason to believe that 9-10wks is a reasonable delivery time, when there are people waiting much longer than that.  If people start confirming that their chips came in, and it looks like the lead time is fixed, I'll change this.

The spreadsheet should be objectively based on the manufacturers advertised lead time. Changing information is personal opinion, highly subjective, misleading, and does not give an accurate portrayal of the companies. A column could be added to describe the likelihood a company will deliver on time, but arbitrarily changing lead times is dishonest, regardless of the intentions.

Additionaly, the prices of companies which are based on opensource designs are not being represented equally, as each is offering a different level of finish. Assembly is not necessarily the same thing from each company. Some are offering parts only, some parts and accessories, some fully assembled units with cases and power supplies.

That being said, thank you for taking the time to work on this.

I actually agree with this.  I tried to keep the table as objective as possible in the manner you describe.  I'm adding BFL back in today or tomorrow for this very reason.  I took them off for more signaling reason's than any belief that they will never deliver.  Unfortunately the only signaling in pushing back Avalon's time is that it makes the people dependent on them look less competitive.  However, I did not move them back for the same reason.  They were moved back because from the information I have found, it seems quite unreasonable to believe that 9-10 weeks is but a remote possibility for the people unlucky enough to not have made an order yet.  Though I really should go change the lead time for companies that are selling their own chips in miners...  That could happen even before the 9-10 week said lead time..  wow I just realized I need to go ask everyone what they think their own personal lead time is based on what their chip order is.  Well again thanks, it's not fair to put everyone together in this way, maybe I can get away with just putting everyone who is selling full miners with a preorder of chips a month out. 
I never intended to say that people who already have preorders wouldn't get their chips until November, better go fix it..

Awesome work! 

Thanks, I'm trying to help people figure out what's going on so we can have more decentralized hashing

[steamboat]

I pushed all the companies requiring Avalon chips back a couple months.  There is no reason to believe that 9-10wks is a reasonable delivery time, when there are people waiting much longer than that.  If people start confirming that their chips came in, and it looks like the lead time is fixed, I'll change this.

The spreadsheet should be objectively based on the manufacturers advertised lead time. Changing information is personal opinion, highly subjective, misleading, and does not give an accurate portrayal of the companies. A column could be added to describe the likelihood a company will deliver on time, but arbitrarily changing lead times is dishonest, regardless of the intentions.

Additionaly, the prices of companies which are based on opensource designs are not being represented equally, as each is offering a different level of finish. Assembly is not necessarily the same thing from each company. Some are offering parts only, some parts and accessories, some fully assembled units with cases and power supplies.

That being said, thank you for taking the time to work on this.

I actually agree with this.  I tried to keep the table as objective as possible in the manner you describe.  I'm adding BFL back in today or tomorrow for this very reason.  I took them off for more signaling reason's than any belief that they will never deliver.  Unfortunately the only signaling in pushing back Avalon's time is that it makes the people dependent on them look less competitive.  However, I did not move them back for the same reason.  They were moved back because from the information I have found, it seems quite unreasonable to believe that 9-10 weeks is but a remote possibility for the people unlucky enough to not have made an order yet.  Though I really should go change the lead time for companies that are selling their own chips in miners...  That could happen even before the 9-10 week said lead time..  wow I just realized I need to go ask everyone what they think their own personal lead time is based on what their chip order is.  Well again thanks, it's not fair to put everyone together in this way, maybe I can get away with just putting everyone who is selling full miners with a preorder of chips a month out. 
I never intended to say that people who already have preorders wouldn't get their chips until November, better go fix it..

Awesome work! 

Thanks, I'm trying to help people figure out what's going on so we can have more decentralized hashing

That is not objective, or based on the figures each company is providing.

Whether or not you believe a company can meet their timelines is irrelevant, and should at the very least be placed in a separate column. Perhaps something like a stated personal belief in the likelihood the company will ship successfully.

Speaking for myself, the lead time for any batch purchased is 9-10 weeks for delivery from avalon, and 3 days for shipping assembled miners. Adding two and a half months to my timeline is an inaccurate representation of what I am offering.

There are plenty of threads on this forum claiming x company will never ship. I believe you are trying to build a spreadsheet that accurately represents the companies in this space and what they are offering, and I thank you for the time you have invested in doing so.

[FCTaiChi]

Right, but the main person that I am watching out for isn't the company, it's the consumer.  In this case I'm having a hard time figuring out how to make it fair.  The company says they will ship in 9-10 weeks.  The company has said a lot of things about times, and none of them have bared out.  In the last few weeks there was a personal announcement about sample chips coming out in a week from Yifu.  It took, what two, three weeks just to ship?  How can I claim that a person will make an ROI of such and such when the proof and general consensus is otherwise?  There have been so many companies that have come on claiming this and that out the first day, people asking why aren't I putting them on the list. 
Here is my method for gathering data:
First look at what the claims made by the company are, sifting for inconsistencies and warning signs.  Then I'll generally check through Coin Canary and talk to Operatr about them.
If the company has been thoroughly tried and the general consensus is that they are at least up to par with basic expectations I'll put them on and watch.  There are companies right now that I feel good about, but I'm checking first.  Just announcing doesn't get you on.  In the same way just saying something doesn't necessarily compel me to post it.  I'm thinking of this as a consumer watch group rather than an aggregation of company data.  Perhaps I'm not being clear on that, there is probably a way to change the title.
Though I could see how splitting columns to something like "Company stats"  "Adjusted stats" or something that wasn't confusing might work... 
So I guess by objective, I meant that I'm trying to take all available data and post it without being attached.  I would like to know what others think is a reasonable estimate on lead time for chips bought today would be.  Maybe November is a bit much, I would be willing to compromise if you have a compelling case.  I thought 3 months was going easy on them since they are WAY past that on their other orders.  How are they going to catch up?

I want the opinion base to be as broad as possible.  This discussion is very valuable to me, and if you or anyone else in the community would like to continue it I'm always open for ideas.

[furuknap]

We have started just a nice discussion here.
I have been thinking about that model used here - constant network growth on each retarget. I don't think it's possible even for a year.
I created a chart with data from this year and +12% on each reatrget. That's exponential growth up to 26,000 Thash by the end of next year (dec 2014)! Just make a chart with all that data and you will see that sharp exponential growth.
I don't see how all that hash power can be added in a year. I see something like linear growth up to 3,000 or maybe 5,000 Thash in that time.

This is a common mistake that opponets of mining investments do; to extrapolate the future based on a limited sample size. Since February, where we had roughly 20TH, we have now grown to 7x and almost touched 140TH just a couple of days ago. In the course of four months, we've increased by 700%, which would mean that a similar growth would put the network at 1PH/s in October, 7PH/s in February 2014, 49PH/s in June 2014, and 280PH/s in December.

Fixing the growth in a number of TH/s per period will also lead to a declining growth over time, so that is equally false, but seems to be a favorite especially when arguing against PMBs. "We've grown 20TH a month the past month, just imagine how huge that number is in a year!" That is actually far better than increasing exponentially...

I wrote a bit about this in my latest article on PMBs as well, which really applies to hardware mining as well: http://coin.furuknap.net/are-perpetual-mining-bonds-scams-not-really/

I'm not sure, obviously, what the future holds, but I'm fairly certain it is neither linear nor exponential.

.b

[ujka]

...
I'm not sure, obviously, what the future holds, but I'm fairly certain it is neither linear nor exponential.

I agree with that. Exp growth is absolutely not sustainable.

[Schrankwand]

They were moved back because from the information I have found, it seems quite unreasonable to believe that 9-10 weeks is but a remote possibility for the people unlucky enough to not have made an order yet.  Though I really should go change the lead time for companies that are selling their own chips in miners...  That could happen even before the 9-10 week said lead time..

For the chips, they are within one week of the designated shipping time of sample chips. You will find the chips aren't the problem in their production right now...

[dmcdad]

They were moved back because from the information I have found, it seems quite unreasonable to believe that 9-10 weeks is but a remote possibility for the people unlucky enough to not have made an order yet.  Though I really should go change the lead time for companies that are selling their own chips in miners...  That could happen even before the 9-10 week said lead time..

For the chips, they are within one week of the designated shipping time of sample chips. You will find the chips aren't the problem in their production right now...

^^^^ THIS ^^^^^

While I don't believe anyone has heard from Avalon regarding the matter, I don't think we can assume that the delays in batch 2 & 3 means that the chips will be delayed by that amount. There are production problems with the batches (as Avalon has stated) that I highly doubt relate to the chips. So then the question is, will Avalon intentionally hold back delivery of 61 batches of chips just because the miner batches 2 & 3 are delayed? Based on the chip sample deliveries (which were just a bit late and not months late), I speculate the answer is No. They will ship the chips relatively soon.

Now that might be wishful thinking on my part (I'm waiting on chips), but I think that argument is just as valid as the one (if not more so) you've used to bump Avalon chip miner dates by several months.

[jimmy3dita]

Thanks, I'm trying to help people figure out what's going on so we can have more decentralized hashing

Looking deeper into your doc (and if you are not using anything different) I'd suggest a quick script for Google Spreadsheet that can be useful for BTC exchange rate.

On one of my recent works I've added the script "MtGox Bitcoin Last Value" from the "tools" dropdown menu, then simply adding the function

Code:

=getBitcoinLastValueFromMtGox()

I got automatically the price in dollars of 1 BTC

[FCTaiChi]

Thanks, I'm trying to help people figure out what's going on so we can have more decentralized hashing

Looking deeper into your doc (and if you are not using anything different) I'd suggest a quick script for Google Spreadsheet that can be useful for BTC exchange rate.

On one of my recent works I've added the script "MtGox Bitcoin Last Value" from the "tools" dropdown menu, then simply adding the function

Code:

=getBitcoinLastValueFromMtGox()

I got automatically the price in dollars of 1 BTC

That's very interesting!  I wonder if it works for other currencies.

So much to respond to here!  I didn't know that the manufacturing wasn't the problem.  But if this is the case what is holding them back?  Is there a link to anyone from the company talking about this, or other proof that they have the chips made.. just can't figure out how to get the home pick-up for UPS to work? 

Wish I had time to do a comprehensive check, I'll get back to this ASAP.

[muyuu]

...
I'm not sure, obviously, what the future holds, but I'm fairly certain it is neither linear nor exponential.

I agree with that. Exp growth is absolutely not sustainable.

It could hold until the next block reward halving. That doesn't seem too infeasible at the moment. However I'm thinking growth will cool off to 5% as the ASIC fevel cools down. A few companies completely failing to deliver will achieve that, by cooling down the demand.

Another variable to add to the scenarios is expected delivery dates. You cannot count amortisation since day 1 when most companies will take a very long time to deliver, except ASICMiner.

[Schrankwand]

While I don't believe anyone has heard from Avalon regarding the matter, I don't think we can assume that the delays in batch 2 & 3 means that the chips will be delayed by that amount.

Sample CHips have shipped.

So much to respond to here!  I didn't know that the manufacturing wasn't the problem.  But if this is the case what is holding them back?  Is there a link to anyone from the company talking about this, or other proof that they have the chips made.. just can't figure out how to get the home pick-up for UPS to work? 

Yifu talked about a lot in the Bitcoin 2013 video. From deliberate sabotage from shipping companies,, pricing, USB controllers that can't handle certain volts, procuring certain other parts...

[.m.]

I would support a guess, that we will see 1 maybe 2 Ph/s in the end of September.
see here - feel free to amend :

https://bitcointalk.org/index.php?topic=228964.msg2452732#msg2452732

[FCTaiChi]

Yifu talked about a lot in the Bitcoin 2013 video. From deliberate sabotage from shipping companies,, pricing, USB controllers that can't handle certain volts, procuring certain other parts...

I'm not sure why a shipping company would do that.  Wouldn't it hurt their reputation?  What incentive is there?
Pricing?
USB controllers and other parts aren't needed to send out chips, unless I'm misunderstanding you.

...
I'm not sure, obviously, what the future holds, but I'm fairly certain it is neither linear nor exponential.

I agree with that. Exp growth is absolutely not sustainable.

It could hold until the next block reward halving. That doesn't seem too infeasible at the moment. However I'm thinking growth will cool off to 5% as the ASIC fevel cools down. A few companies completely failing to deliver will achieve that, by cooling down the demand.

Another variable to add to the scenarios is expected delivery dates. You cannot count amortisation since day 1 when most companies will take a very long time to deliver, except ASICMiner.

I agree that the difficulty will continue to have some crazy slopes and rises.  The average difficulty of all total rises has been about 17%.  It's handy to say 17%, but it discounts all of the major changes that have happened.  Basically this is what I'm doing with the 1 year estimate.  There are situations where a 12% average may end up actually giving the ROI of a 15%, if the growth is weighted toward the beginning of the cycle.  That's why I added the higher percentage columns, it looks like we will have a fairly low growth cycle compared to current estimates, but it could be heavily weighted toward the near future.

[TheSwede75]

Maybe you should change BFL 5 gh/s to 'YES' on shipping. They have shipped some 1k units in the last two weeks. If Avalon is shipping so is BFL.

[FCTaiChi]

Maybe you should change BFL 5 gh/s to 'YES' on shipping. They have shipped some 1k units in the last two weeks. If Avalon is shipping so is BFL.

They might ship some product, but are they currently shipping anything that you can buy in their store?  Being that this is about lead time I have to take into account only the miners that are ordered today.